🎬 Upload your results! 💡 They give me new ideas on what to improve. The workflow is inside the mp4 — just drop it in the gallery 📥 Failures welcome too 🐛🔧# v1.1 update ## Form fields (new version)- **Version name:** v1.1 — 🖼️ Image to Video + ✨ Refine- **Files:** <code>H3_LongTake_workflows_v1.1.zip</code> (4 workflows + README)- **Gallery (this version):** <code>08_i2v.mp4</code> (image to video + refine from one picture, 15 s, source picture on the left)- **Extra tags:** image to video, hires fix## Version description (paste into the version's "About this version" box)🚀 **v1.1 — Image to Video + Refine** (needs node pack **1.1.1+** → update from ComfyUI-Manager)🖼️ **New workflow <code>H3_LongTake_i2v</code> — a long video from a single picture.** <code>start_image</code> is the first frame,<code>prompts</code> holds **one block per 5-second clip**, separated by a <code>---</code> line (block 1 = 0–5 s, block 2 = 5–10 s…,the last one repeats). One clip in memory whatever the length; 🔊 the audio is generated by H3 per clip.🧑‍🦰 **Identity references against drift** (ref2va model): the picture as <code>&lt;Picture 1&gt;</code> and a close-up of the face as<code>&lt;Picture 2&gt;</code>. 📏 Measured on 30 s: without a reference the face becomes another person within 20 s (similarity0.27), with the face reference it stays at 0.60.✨ **New node Refine (second pass) — video "hires fix"** on any finished project (video to video or image to video):0.5 MP → 1.0 MP with <code>denoise 0.25</code>, real skin and hair detail, and with the picture + face connected it puts thepicture's face back on the whole video even where the first pass lost it (📏 0.27 → 0.62). Output in <code>&lt;project&gt;/hr</code>,assembled by the Stitch like any project. ⏱️ About 6 minutes per clip on 16 GB.📝 **README note inside every workflow** (what it does, settings, steps) — and tidier boxes and titles.📦 **Extra models for the i2v workflow:** <code>minimax_h3_fl2va_pruned_fp8_scaled.safetensors</code> +<code>minimax_h3_fl2v_lightx2v_turbo_4step_v0.1_comfy.safetensors</code> (no references), or the ref2va model + Turbo ref2v ofthe other workflows (with references).📖 Full docs, measurements and prompting guide: <a target="_blank" rel="ugc" href="https://github.com/mark9009/ComfyUI-H3-LongTake">https://github.com/mark9009/ComfyUI-H3-LongTake</a>## One line to add at the top of the main description (optional)🆕 **v1.1:** 🖼️ Image to Video workflow (a long video from one picture, one prompt block per 5 s) and ✨ Refinesecond pass at 1.0 MP — see the version notes.Three ComfyUI workflows for MiniMax H3 that turn a source video of any length into a new video, clip by clip,on a single 16 GB GPU. They need the custom node pack ComfyUI-H3-LongTake (MIT, no extra Python dependencies):👉 <a target="_blank" rel="ugc" href="https://github.com/mark9009/ComfyUI-H3-LongTake">https://github.com/mark9009/ComfyUI-H3-LongTake</a> — install, models list, tested use cases with the exact prompts, prompting guide.What the node does- Slices the source video into 124-frame clips (5.2 s) that overlap by 5 frames, decodes only each slice with ffmpeg, and renders them one after the other.- Every clip gets the previous clip's latent tail as an H3 keyframe, so the seam is continuous (measured 0.96 continuity) and the video follows the source with no drift over minutes of material.- Resumable projects: stop ComfyUI, queue again, it continues from the first missing clip. Redo a single clip <code>redo_one</code>) or everything from clip N <code>redo_from</code>) with a new seed while the seams stay.- <code>seam_match = color</code> removes the colour/exposure jump between clips (ΔE 5.8 → 0.9).- The Stitch node concatenates the clips without re-encoding, puts the original audio back and embeds the workflow in the mp4 (drag the result into ComfyUI to reopen the project).The three workflows1. H3_LongTake_character_swap — a picture's character performs the whole video <code>source_role = reference</code>, Ref2VA). Identity from <code>&lt;Picture 1&gt;</code>, motion, timing, camera and room from the video. Works with photoreal and anime references; the video's room is kept even when the picture has its own background (say so in the prompt).2. H3_LongTake_example — the same, with the structured H3 prompt (subject definitions / retention analysis) for motion + identity transfer.3. H3_LongTake_style — style transfer and retexture with the StyleTransfer LoRA by NRDX (<a target="_blank" rel="ugc" href="https://civarchive.com/models/2932297">https://civarchive.com/models/2932297</a>): the video becomes a guide latent <code>source_role = guide</code>), motion and framing are reproduced frame by frame (adherence 0.93), the style comes from text (Van Gogh, Ghibli, Simpsons, GTA…) or from a style swatch as <code>&lt;Picture 1&gt;</code>. <code>retexture:</code> changes the material/colour of one garment and nothing else.Models (all from the official Comfy-Org MiniMax-H3 release unless noted)- <code>minimax_h3_ref2va_pruned_fp8_scaled.safetensors</code> (diffusion_models)- a MiniMax H3 Qwen3-VL-32B text encoder build (int8 or nvfp4; ~25 GB, kept in system RAM)- <code>minimax_h3_video_vae_fp16.safetensors</code>, <code>minimax_h3_audio_vae_fp32.safetensors</code> (vae)- <code>minimax_h3_ref2v_turbo_4step_v0.1_comfyui_bf16.safetensors</code> (loras) — 4 steps- optional: <code>minimax_h3_style_transfer_v1.0_r64.safetensors</code> by NRDX (loras) — style workflow onlyQuick start1. Load a workflow, put your video in <code>ComfyUI/input/</code> and pick it in <code>source_file</code>; connect your picture to <code>ref_image_1</code>.2. <code>dry_run = true</code> → the node prints the slicing plan.3. <code>dry_run = false</code>, <code>max_clips = 2</code> → check identity/style and the seam in the preview.4. <code>max_clips = 0</code>, <code>mode = continue</code> → the whole video. Unmute the Stitch node (Ctrl+M) when all clips are done.Timing on a 16 GB GPU at 544×960: about 4:40 per 5-second clip with 4 Turbo steps (~1 minute of video per hour).Prompting rules that came out of the tests- Named styles work from text alone: 3–5 visual attributes + "Keep the original room, background, framing and motion exactly as in the video; do not add any new scenery." Never a style name and <code>&lt;Picture 1&gt;</code> in the same sentence.- Style pictures without faces or scenery (crop a swatch): a face transfers the identity, a skyline replaces the room.- Character swap: describe the character in words as well as <code>&lt;Picture 1&gt;</code>, and forbid the picture's background explicitly. A picture framed like the video keeps the motion in sync.- <code>retexture:</code> names the exact garment ("the dress", not "the outfit").Examples in the gallery (source | result): character swap from a photoreal picture, Van Gogh, Studio Ghibli, GTA from a swatch, The Simpsons on a 16:9 press conference, pop-art from a halftone swatch, gold-armour retexture. Source clips are free Pexels videos and a film excerpt; the character reference is an AI-generated picture.Credits: motion-context idea by NikoDemon80 and the Banodoco MiniMax H3 thread; StyleTransfer LoRA by NRDX; MiniMax H3 core nodes by Comfy-Org. Node pack and workflows: mark9009.

MiniMax H3 LongTake — long videos clip by clip | character swap, style transfer, retexture

# Civitai — v1.3 update only## Form fields (new version)- Version name: v1.3 — ⚡ latent upscaler in Refine + 🧑‍🎨 character sheets- Files: <code>H3_LongTake_workflows_v1.3.zip</code> (4 workflows + README)- Gallery (this version): reuse <code>08_i2v.mp4</code> or add a new Refine clip if you have one- Extra tags: latent upscale## Version description (paste into the version's "About this version" box)🚀 v1.3 — faster, sharper Refine and character sheets (needs node pack 1.3.0+ → update from ComfyUI-Manager)⚡ Refine with the H3 latent upscaler. The second pass now enlarges the latent instead of decoding,resizing and re-encoding the video <code>upscale = latent_model</code>, the default). 📏 Measured per clip, samereferences and seed: 253 s instead of 375 s, face sharpness 222 vs 85, same identity, hands +60 %.📦 Install the Comfyui_Minimax_h3_latent_Upscaler node pack and put<code>minimax_h3_latent_upscaler_3d_fp16.safetensors</code> in <code>models/latent_upscale_models</code>. Without them the Refinefalls back to the old pixel upscale and tells you in its report — nothing breaks.🎞️ Refine from edited chunks <code>frames_from = mp4</code>): refine clips you already touched up in pixels (a facefix, a colour grade) instead of the raw latent.🧑‍🎨 Character sheets (from 1.2.0, never announced here): <code>character_sheet</code> on Image to Video and Refine, asheet with several views of the character as an extra reference. 📏 Use it on top of the picture + facereferences: alone it is worse, together it raises the identity floor (min 0.16 → 0.26 on a hard 15 s test).🧠 Newer ComfyUI (comfy-aimdo dynamic VRAM) fixed: the second run in the same session no longer slows downuntil it hangs — the nodes free VRAM before each heavy phase.🛡️ Safer dry runs: with the Render in <code>dry_run</code>, Refine and Stitch stop with a message instead ofre-assembling an older project with the same name. <code>project_name</code> is greyed out while <code>project_dir</code> is linked.🔧 Fixed in the i2v workflow: the Stitch's <code>audio_file</code> held an Italian value that failed validation onceyou unmuted it. The Refine node now shows the new <code>upscale</code> / <code>frames_from</code> options and its README noteexplains the latent upscaler.📖 Full docs, measurements and prompting guide: <a target="_blank" rel="ugc" href="https://github.com/mark9009/ComfyUI-H3-LongTake">https://github.com/mark9009/ComfyUI-H3-LongTake</a>## One line to add at the top of the main description (optional)🆕 v1.3: ⚡ Refine with the H3 latent upscaler (a third faster, much sharper faces) and 🧑‍🎨 charactersheets — see the version notes.

Files with SHA256: ecc4ec13f41ac775dd64c4743974194b54fa9cd6496ffdc947f2b101f015a22e